Chemical properties of polyamines with relevance to the biomineralization of silica

Article abstract of DOI:10.1039/b310201g

Polyamines mimicking substances which occur naturally in biosilicas have been synthesized and show an accelerating effect on silica condensation, which depends on the chemical nature, the architecture (linear or branched), and the degree of polymerization.

Full text of DOI:10.1039/b310201g

Chemical properties of polyamines with relevance to the
biomineralization of silica†
Henning Menzel,*a Sandra Horstmann,a Peter Behrens,*b Petra Bärnreuther,b Ilka Kruegerb and
Michael Jahnsb
a Institut für Technische Chemie, Technische Universität Braunschweig, Hans-Sommer-Str. 10, 38106
Braunschweig, Germany. E-mail: h.menzel@tu-bs.de; Fax: 49 531 391 5357; Tel: 49 531 391 5361
b Institut für Anorganische Chemie, Universität Hannover, Callinstr. 9, 30167 Hannover, Germany.
E-mail: peter.behrens@mbox.acb.uni-hannover.de; Fax: 49 511 762 3006; Tel: 49 511 762 3660
Received (in Cambridge, UK) 22nd August 2003, Accepted 23rd October 2003
First published as an Advance Article on the web 7th November 2003
Polyamines mimicking substances which occur naturally in
biosilicas have been synthesized and show an accelerating
effect on silica condensation, which depends on the chemical
nature, the architecture (linear or branched), and the degree
of polymerization.
prepolymers are subsequently either hydrolysed by aqueous
sodium hydroxide, yielding linear PEI or PPI, respectively, or
are subjected to a combined hydrolysis/Leukart–Wallach reac-
tion, yielding the methylated variants, PMEI or PMPI, re-
spectively. Both polymer-analogous reactions do not change Pn.
The hydrolysis is almost complete as no residual acetyl groups
1
Important bioorganic components taking part in the bio-
mineralization process of silica have been identified recently.1
For example, polysaccharides1 and proteins (silicatein from
sponges,2 silaffin from diatoms3,4) have been isolated from
biosilica minerals. For the biomineralization of silica in
diatoms, also simple polyamines as polypropylene imine
derivatives play an important role.5 Sumper has developed a
model to explain the pattern formation in diatom shells by phase
separations occurring in silica solutions of these polyamines.6
We describe the synthesis of polyamines modelled on the
naturally occurring substances as well as their acid-base
behaviour and their influence on the kinetics of silica con-
densation.
In order to model the polyamines found in diatoms, which are
all linear and have a relatively low degree of polymerization Pn
of 10 to 20 monomer units, we have synthesized polyamines by
ring-opening polymerization of oxazolines or 1,3-oxazines,7
respectively, and subsequent hydrolysis.
Linear polyethylene imine (PEI) was prepared by polymer-
isation of methyloxazoline (Scheme 1). The prepolymer
polyacetylethylene imine has a molecular weight of approx-
imately 830 g mol21 corresponding to a degree of polymer-
isation Pn = 8–9. Linear polypropylene imine (PPI) can be
prepared with essentially the same chemistry, with 1,3-oxazine8
used as monomer (Scheme 2). The corresponding prepolymer
has a Pn of 12–13 (determined by NMR spectroscopy). The
can be detected by H NMR. The degree of methylation was
verified by means of NMR spectroscopy to be larger than 99%
for PMEI as well as for PMPI. 1,5,9,14,18,22-hexaazadocosane
is prepared by Michael addition of acrylonitrile to spermine.9
It is expected that the pH of a polyamine-silica solution
strongly influences the condensation behaviour. We therefore
characterized the acid-base behaviour of the polyamines used in
this study by collecting titration curves.10,11
Furthermore, we found that the solutions of polyamines used
in the condensation experiments (see below) always show an
initial pH of 4.3 to 4.8. The system polyamine-silica thus is
strongly buffering the solution. It is noteworthy that this
buffering pH is close to the pH found in silica deposition
vesicles.12
There are significant differences in the titration curves for the
various polyamines (Fig. 1). Linear PEI shows two distinct
steps in the titration curve. At high pH, the majority of the
amino groups is neutral; protons start to bind to the chain
independently up to a situation where approximately one half of
the amino groups are protonated. At this point, a polyamine
chain tends to minimize its free energy by protonation of only
Scheme 1
Scheme 2
† Electronic Supplementary Information (ESI) available: detailed informa-
tion about the synthesis and the kinetic investigations. See http://
www.rsc.org/suppdata/cc/b3/b310201g/
Fig. 1 Titration curves for the polyamines (c = 8.6 · 1024 mol L21 amino
groups); top: ethylene imines, bottom: propylene imines.
2994
CHEM. COMMUN., 2003, 2994–2995
This journal is © The Royal Society of Chemistry 2003

View Article Online
every second amino group, to avoid nearest neighbor inter-
actions of ionic sites. When then an additional amine site is to
be protonated, it has two nearest neighbors which are already
protonated and two direct pair interactions have to be overcome.
The pKb for the remaining amino groups therefore is sig-
nificantly reduced.11 The branched PEI (purchased from
Aldrich) shows a smooth titration curve without a plateau,
because there is a large number of different amino groups,
which all have different connections to their next neighbors due
to the irregular branching scheme. Therefore, for branched PEI
no distinct pKb but an almost continuous distribution of pKb
values for all the different amino groups is expected.10,11 PMEI
shows a titration curve similar to that of linear PEI. There is only
a small shift, indicating a higher basicity of the PMEI.
For the polypropylene derivatives, i.e. PPI, PMPI and
hexaazadocosane, the respective titration curves are very
similar. All show two distinct steps as expected for linear
polyamines.11 The shift of the titration curves reflects the
differences in the basicity between PPI and PMPI. 1,3-diamino-
propane shows a significantly different titration curve, because
the amino groups in this compound are primary amino groups,
while all other compounds have mostly secondary amino
groups. Furthermore, in 1,3-diaminopropane there is only one
neighbouring amine group present, while in the polymeric and
oligomeric amines, most amino groups have two neighbours.
Polyamines have been shown to influence the condensation
of silicic acid13 and the flocculation of the primary particles to
form larger aggregates.10 We investigated the kinetics of the
silicic acid condensation reaction in the presence of different
polyamines via the molybdate method, which detects mono-
meric and dimeric silicic acid.1,14 Polyamine solutions were
prepared by dissolving the amine in 1025 M hydrochloric acid
so that the amine concentration was 10 mg L21. A fresh solution
of silicic acid was then mixed with the solution containing the
amine. The final pH of the reaction mixture was between 4.3
and 4.8 for the polyamines and 6 for diaminoethane and
1,3-diaminopropane.† The results for the polyamines with
ethyleneimine-building blocks are shown in Fig. 2.
A strong accelerating effect is found for the 1,2-diaminopro-
pane and linear PEI having a molecular weight comparable to
that of the polyamines isolated from diatoms.5 The effect is less
pronounced for the high molecular mass branched PEI and
methylated PMEI.
The results for the polyamines with propyleneimine building
blocks show qualitatively similar results (Fig. 3). The strongest
acceleration is found for the 1,3-diaminopropane, followed by
hexaazodocosane and the linear PPI. The acceleration effect
again is much smaller for the methylated variant (PMPI). The
difference between hexaazodocosane, which can be regarded as
a PPI with Pn = 6 and the linear PPI (Pn = 12–13) indicates that
the chain length influences the accelerating effect.
Fig. 3 Condensation kinetics of silicic acid in solutions containing 10 mg
L21 polyamine: Amines with propylene imine building blocks. Plot of the
normalized absorption of the molybdatosilicate as a function of time.
PMEI, respectively). However, the chain lengths of the
polypropylene amines are larger than those of the polyethylene
derivatives, and in view of the results described above, this
could well be the reason for the observed differences.
In summary, polyamines modelled on naturally occurring
substances have been synthesised and investigated with respect
to their effect on silicic acid condensation. The accelerating
effect depends on the chemical nature (polyethylene or
polypropylene imine, degree of methylation), the architecture
(linear or branched) and the degree of polymerization of the
polyamine. Commercially available PEI with high molecular
mass and a branched architecture, which was used in other
studies,12 is not a good model for the biological polyamines, as
its chemical behaviour is significantly different from that of
linear PEI and PPI. Among the polyamines investigated, PMPI
is closest to the naturally occurring polyamines. Its acceleration
effect is smaller than that of non-methylated PPI or PEI.
Obviously, nature does not use the system accelerating the silica
condensation most efficiently, but rather uses a slower system,
probably to provide sufficient time for the formation of the
intricate macrostructures.
The authors acknowledge support from the DFG within the
Special Research Programme “Prinzipien der Biomineralisa-
tion” and by the Fonds der Chemischen Industrie.
Notes and references
1 C. C. Perry and T. Keeling-Tucker, J. Biol. Inorg. Chem., 2000, 5,
537.
2 K. Shimuzu, J. Cha, G. D. Stucky and D. E. Morse, Proc. Natl. Acad.Sci.
U.S.A., 1998, 95, 6234.
3 N. Kröger, R. Deutzmann and M. Sumper, J. Biol. Chem., 2001, 276,
26066.
Linear PPI (linear PMPI) appears to be less effective in
accelerating the silicic acid condensation than linear PEI (linear
4 N. Kröger, S. Lorenz, E. Brunner and M. Sumper, Science, 2002, 298,
584.
5 N. Kröger, R. Deutzmann, C. Bergsdorf and M. Sumper, Proc. Nat.
Acad. Sci. U.S.A., 2000, 97, 14133.
6 M. Sumper, Science, 2002, 295, 2430.
7 S. Kobayashi and T. Saegusa, in Ring-opening Polymerization, K. J.
Ivin and T. Saegusa Eds., Elsevier, London, 1984, pp. 761.
8 T. Saegusa, Y. Nagura and S. Kobayashi, Macromolecules, 1973, 6,
495.
9 B. Dietrich, M. W. Hosseini, J.-M. Lehn and R. B. Sessions, Helv. Chim.
Acta, 1983, 66, 1262.
10 C. M. Lindquist and R. A. Stratton, J. Coll. Interf. Sci., 1976, 555,
45.
11 G. J. M. Koper, R. C. van Duijenbode, D. D. P. W. Stam, U. Steuerle and
M. Borkovec, Macromolecules, 2003, 36, 2500.
12 E. G. Vrieling, T. P. M. Beelen, R. A. van Santen and W. W. C. Gieskes,
Angew. Chem., Int. Ed. Engl., 2002, 41, 1543.
13 T. Mizutani, H. Nagase, N. Fujiwara and H. Ogoshi, Bull. Chem. Soc.
Jpn., 1998, 71, 2017.
Fig. 2 Condensation kinetics of silicic acid in solutions containing 10 mg
L21 polyamine: Amines with ethylene imine building blocks. Plot of the
normalized absorption of the molybdatosilicate as a function of time.
14 B. G. Alexander, J. Am. Chem. Soc., 1953, 75, 2887.
CHEM. COMMUN., 2003, 2994–2995
2995